The cost of guessing
Counting the work that came back
A line per return, four fields, thirty seconds. What to record, what the record is for, and the two ways it goes wrong.
Decisions this page is about, and where the line actually was
Changing the wording on a customer-facing error message
4 → 1It had been yours all along
Refunding a customer under fifty pounds
3 → 1A written policy you had never been shown
Telling a customer a date
3 → 6Dates are committed by one person here
Reordering your own week
4 → 1You asked twice before believing the answer
- 1Decide it alone and say nothing
- 2Decide it alone and mention it afterwards
- 3Check with one colleague first
- 4Check with the person who assigned it
- 5Needs a decision one level up
- 6Cannot be decided here at all
four decisions · three of them you had more room than you thought · one gave you less · the median gap was three steps on the scale
Four of the nine decisions counted on this site.
Where ownership crosses several roles, teams can review the official platform to make time and recurring bottlenecks visible enough for a useful operational conversation.
A separate professional source for this subject is International Labour Organization.
Keeping this record takes about thirty seconds per entry and perhaps six entries a quarter. It is the only measurement on this site that produces a number somebody else will believe.
The four fields
What the work was. Enough to recognise it later. Not a ticket number — a phrase.
How far it got. The stage it had reached when it came back: brief, plan, build, review, shipped. Five is plenty; more is false precision.
What it went back to. The same scale. The distance between the two is the cost.
Why. One clause, written the day it happened, in the plainest available words. "Nobody had said who the audience was." "A rule existed that nobody mentioned."
That is the whole record. Six lines by the end of a quarter.
Writing the reason honestly
The reason field is the only one requiring judgement, and it is where the record becomes either useful or worthless.
Write what was missing, not who was at fault. "The constraint on downtime was never stated" is a fact about information. "X forgot to tell me about the downtime" is the same fact framed as a grievance, and a record written that way cannot be shown to anybody, which means it cannot do its job.
Write it the same day. A week later the reason will have been smoothed into "we misunderstood the requirements", which is true, contains nothing, and matches every other entry.
And record the returns that were your own fault. A record containing only other people's failures is an argument, not a measurement, and the first person to read it will notice.
What the record is for
Not for performance conversations, and this needs saying explicitly because it is the obvious misuse.
It is for one question, asked at the end of a quarter: what would have had to be said at the start, for these six not to have happened?
On the record kept for this site the answer was unambiguous — four of six were a missing constraint or a missing purpose, both of which fit in a sentence. That is an actionable finding. "We should communicate better" is not.
The two ways it goes wrong
It becomes a defect log. Somebody starts recording every review comment, the count reaches forty, and the signal disappears into ordinary review activity. Record returns that cross more than one stage; single-stage returns are review working.
It becomes evidence. The moment the record is used to establish that somebody is difficult to work with, it stops being written honestly, and within a quarter every reason field says "requirements changed". There is no technical protection against this. The only protection is that whoever keeps it declines to use it that way, every time, including when it would be convenient.
A worked example of the record
Six entries, as they were actually written during the quarter counted here, with nothing added afterwards:
- Import rewrite — review to plan — the constraint on downtime was never stated.
- Quarterly numbers — shipped to brief — nobody had said who the audience was.
- Signup flow — review to build — a rule existed that nobody mentioned.
- Migration guide — review to plan — it was written for the wrong reader.
- Reporting page spec — build to brief — the purpose turned out to be a different one.
- Vendor handover — review to build — a previous attempt had failed for a reason nobody passed on.
Six lines. Perhaps ninety words in total, written over thirteen weeks.
What makes them usable is that each reason is a specific absence rather than a category. "Requirements changed" would have fitted all six and told nobody anything. "Nobody had said who the audience was" identifies a slot, a moment and a sentence that would have prevented it.
Reading them together, the finding is immediate and did not require analysis: four of six name a constraint or a purpose, both of which fit in one line of a handover message.
Two quarters in
The second quarter's record is where the value is, because it can be compared with the first.
What you are looking for is not a lower count. Counts move with how much work there was and how experimental it was. You are looking at the reason column: whether the same cause is still producing returns.
If "missing purpose" appears three times in quarter one and once in quarter two, something changed and you can say what. If it appears three times in both, nothing was fixed and the record has told you so in a form that is hard to argue with.